The study of deep learning and neural networks is of growing importance in today's world, as artificial intelligence (AI) increasingly impacts many aspects of our lives - from autonomous vehicles and medical diagnostics to financial forecasting and robotics. At the core of these technologies lie mathematical methods that enable machines to learn from vast amounts of data. In this course, we delve into the theoretical aspects of deep learning, exploring its mathematical foundations, training algorithms, and network structures, while also addressing related problems such as convergence, stability, and interpretability.
We combine theory with practice by implementing neural networks ourselves, using the PyTorch framework - a powerful tool that lets us efficiently construct, train, and deploy deep models. During the practical sessions, we apply these methods to solve real-world problems, such as image recognition with convolutional neural networks and handwritten digit recognition using the MNIST dataset. This approach prepares students to understand the fundamental mathematical aspects of deep learning and to gain practical experience in implementing and testing AI-driven applications.
Link zum Modulhandbuch Mathematik
Weitere Informationen (pdf-Datei)
Modul MAT-4xx
Students are recommended to have their own laptop with Python environments installed, preferably version 3 or higher, and a Python editor such as Jupyter Notebook, Visual Studio Code, or a similar tool. It is also desirable to have Git-related software installed